Reference #

Authoritative listing of Litmus's public surface, organized by the same category axis as concepts and how-to. Each entry links to the page that exhaustively documents one boundary.

Overview #

The foundational pages — pytest mechanics every Litmus test uses, plus the inventory of AI / MCP skills the platform ships.

• pytest-native — how a Litmus test uses pytest's own collection / fixtures / markers / conftest.py / CLI flags. Nothing Litmus-specific. The baseline every other page builds on.
• Skills — the workflows, sub-agent templates, slash commands, and MCP prompts that ship with Litmus for AI-assisted authoring.

Configuration #

The YAML schemas every entity validates against, plus the catalog (its own thing because the catalog is shared across projects).

• Configuration — litmus.yaml, station YAML, fixture YAML, product YAML, sidecar YAML, profile YAML — all schemas the fixtures and markers read from (generated from the Pydantic models).
• Catalog → schema — every field in a catalog/<vendor>/<model>.yaml entry, the rules, the "what goes WHERE" decision tree.
• Catalog → cookbook — worked recipes for the recurring datasheet shapes (accuracy bands, dual-unit values, shared controls, conditional attributes, etc.).

pytest plugin #

The fixtures and markers the bundled pytest plugin contributes on top of stock pytest.

• Fixtures — all 20 fixtures the plugin contributes on top of pytest's, with signatures, scopes, and per-fixture examples.
• Markers — the seven @pytest.mark.litmus_* decorators the plugin registers and their 1:1 sidecar equivalents.

Data #

The shapes the system writes. If you're reading parquet, the event log, or any export — these pages describe exactly what's in them.

• Models — every public Pydantic model + ERD of how they reference each other (generated).
• Event types — every typed event payload the runtime emits (generated).
• Parquet schema — every column in the run parquet, the record_type discriminator, how retries land.
• Output formats — what litmus show -f <fmt> and litmus export produce for HTML / PDF / JSON / CSV / STDF / HDF5 / TDMS / MDF4 / ATML.
• Query API — RunsQuery, StepsQuery, MeasurementsQuery. The public read path the UI and HTTP API both use (generated).

Runtime #

The interactive and programmatic surfaces — for LabVIEW, TestStand, scripts, dashboards, AI agents.

• LitmusClient — Python client that submits test runs (no pytest required). Suits LabVIEW / TestStand bridges.
• connect() — interactive instrument access for scripts, notebooks, the operator UI. Returns a StationConnection with the full event-log / channel-store / instrument-pool surface.
• HTTP & MCP API — REST endpoints exposed by litmus serve, plus the twelve MCP tools (generated). Same shapes either way.

Operator UI (litmus serve) #

Per-screen reference for the browser surface. For orientation, see the Tour of the Operator UI.

• Dashboard — /
• Launch Test — /launch
• Live monitor — /live/{run_id}
• Results — list — /results
• Results — detail — /results/{run_id}
• Metrics — /metrics
• Measurements — /explore
• Events — /events
• Channels — list — /channels
• Channels — detail — /channels/{channel}
• System Designer — /designer
• Stations — /stations
• Products — /products
• Fixtures — /fixtures
• Instruments — /instruments
• Tests — /tests

Command line #

• CLI reference — every litmus <command> and its flags (generated).